The emergence of modern geospatial techniques has brought a considerable change in data capturing, processing, and visualization tools. Digital elevation models (DEMs) are a 3D mathematical representation of the earth’s surface. Due to enormous advancement in geographic information system (GIS) tools and technologies, DEMs have a wide range of applications in natural resources management, environmental science, and engineering. In general, primary topographic variables extracted from DEMs are frequently less useful than secondary ones, e.g., surface roughness. Therefore, the precise values of secondary terrain features need to be derived from base DEMs. Generally, topographic roughness is an essential geomorphological variable that is a vital issue in geoscience. This research presents a comparison of deterministic interpolation methods, inverse distance weighted (IDW) and a natural neighbor (NN), with geostatistical technique, ordinary kriging (OK), to investigate the influence of generated DEM on quantifying terrain roughness relied on synthetic data and applying the zonal statistics tool. At the same time, a quantitative and qualitative approach is adopted to assessing the results through statistical methods and visual representations. The findings indicated that roughness values' behavior is broadly related to the quality of the built DEM. The NN approach yielded the greatest DEM accuracy, standard deviation, from a quantitative perspective (±0.930 m) than IDW (±3.748 m) and OK (SD = ±5.544 m) methods.
Read full abstract